A
SpikeProblem is brief experiment to validate that user really has the problem we hypothesize they do, before we invest in a production implementation. It is the user-facing counterpart to the
SpikeSolution. The idea of
SpikeProblem or
SpikeUserStory is inspired by the
LeanStartup work of
EricRies and customer-development work of
SteveBlank. Ries and Blank recommend quickly validating that customers really want what we propose to build, as fast as possible, as a way to minimize the risk and waste of building the wrong features/products..
SpikeProblem's can utilize simple
LandingPage's,
FeatureFakes, A_B_Testing, etc. They can also include non-software approaches including customer interviews, observation, paper prototypes, etc. It is often helpful to have the development team participating in this type of customer discovery, and they can often instrument the software to gather validation information. This type of validation can occur on the scale of minutes, hours, and days, usually less than a typical Agile iteration period. Often many such experiments are run in parallel, with different customers. Because speed of learning is the highest priority, many of these
SpikeProblem's can be completed with minimal time and less-than-production quality. The team needs to use engineering and business judgement to strike the right balance between a
SpikeProblem and the on-going production code. --
KelleyHarris